LLM (Large Language Model)
An LLM (Large Language Model) is an AI model trained on very large amounts of text so it can read, summarise, answer questions and write. It is the technology behind ChatGPT, Claude and Gemini, and it can be built into your own software to handle documents, emails and customer queries.
Key Facts
| Good at | Summarising, extracting fields from documents, drafting, classifying, answering from supplied material |
|---|---|
| Weak at | Exact calculations, facts it was not given, anything needing guaranteed accuracy without checks |
| Deployment options | Hosted APIs (OpenAI, Anthropic, Google) or open-weight models on your own servers |
| Cost driver | Volume of text processed, measured in tokens |
How businesses use LLMs
- Reading invoices, purchase orders and contracts into structured data.
- Answering staff or customer questions from company documents.
- Drafting replies, proposals and summaries for a person to review.
Making LLM output reliable
Give the model the right source material (see RAG), constrain what it may answer, validate outputs, and route uncertain cases to a person. See controlling hallucination.
Frequently Asked Questions
Will an LLM learn from my company data?
Not unless you train or fine-tune it. Business API plans from major providers generally do not use your inputs for training; check the terms of the plan you use.
Can I run an LLM on my own servers?
Yes, with open-weight models. It suits sensitive data or high volumes but needs GPU hardware and maintenance.
Related Glossary
Need help implementing this in your business?
Turbo Bytes Consulting helps businesses streamline operations and build custom software architectures that scale without chaos.